Loading
Conformal prediction wraps any model so that, instead of one answer, it returns a set or interval guaranteed to contain the truth a chosen fraction of the time (say 90 percent), assuming new patients resemble the calibration patients.
Conformal prediction produces statistically valid prediction regions for any underlying point predictor, assuming only exchangeability of the data, by computing nonconformity scores on labelled calibration data (Wikipedia). It turns a subtype classifier into one that may say luminal A or luminal B when unsure, with a coverage guarantee, and it makes the exchangeability assumption explicit, which domain shift violates. Angelopoulos and Bates give the standard introduction.
Shares Out-of-distribution detection (Mahalanobis guard), Uncertainty quantification and confidence gates, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Calibration: reliability diagrams and the Brier score, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Uncertainty quantification and confidence gates, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Calibration: reliability diagrams and the Brier score, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Out-of-distribution detection (Mahalanobis guard), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Calibration: reliability diagrams and the Brier score, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Out-of-distribution detection (Mahalanobis guard), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Calibration: reliability diagrams and the Brier score, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.